analogue of what we call self-consciousness in ourselves. But whether there be one Self or many selves, and, if there be both, what is the relation between the One and the many-these are questions of metaphysics or ontology, not to be settled out of hand by the perfectly general result to which the theory of knowledge leads us.

Unquestionably the results of the epistemological investigation must have an important bearing upon the metaphysical problem; but the office of the theory of knowledge must, in the main, be negative or indirect, ruling out certain solutions as inadmissible rather than itself supplying us with a ready-made solution. In a word, the theory of knowledge, even in its amended form, must maintain the critical attitude at first assigned to it by Kant. Though we may disagree with many of the arguments by which he supports his position, it cannot, I think, be doubted that Kant was methodically correct in the view he took of his own inquiry. There is nothing in it, as I conceive, to preclude us from the attempt to construct a metaphysical system; but it cannot stand itself as a dogmatic theory.

Kant himself, it is almost superfluous to point

out, would never have acquiesced in the deductions which his Neo-Kantian followers have drawn from his premisses. Nothing, of course was further from his thoughts than an identification of the transcendental Ego with the divine self-consciousness, as is sufficiently. proved by his constant references to the latter as a perceptive, that is, a non-discursive understanding, the very possibility of which we are unable to comprehend.1 But Kant further refuses to recognise the transcendental Ego as constituting the real self even of the individual human knower. This is, in fact, the text of his

1 As if anticipating that the attempt would be made to represent the difference between the human consciousness and the divine as essentially one of degree, Kant expressly declared himself on this point in an important letter to Marcus Herz in 1789. It will be found, he says, 'that we cannot assume the human understanding to be specifically the same as the divine, and only distinguished from it by limitation—¿.e., in degree. The human understanding is not, like the divine, a faculty of immediate perception, but one of thought, which, if it is to produce knowledge, requires alongside of it— -or rather requires as its material-a second quite different faculty, a faculty or receptivity of perception."-Werke, viii. 719. As further emphasising the complete distinction existing in Kant's mind between the consciousness of the individual and the divine selfconsciousness, reference need only be made to the thoroughly transcendent conception of God with which the Kantian ethics end—a being apart, whose function it is to mete out happiness in accordance with desert.

whole contention in the well-known argument headed "The Paralogism of Pure Reason." Kant is there attacking the old metaphysical psychology for reasoning, not indeed to the same conclusion, but on precisely similar lines to those on which the Neo-Kantian proof of the universal Self has been seen to run. The metaphysical psychologists also started with the abstract Ego, which forms the presupposition of knowledge; and as this unity of consciousness is one, eternal (or out of time), and indivisible, they proceeded to prove by its means the necessary immortality of the human soul. This is the Paralogism which Kant attacks, and in the course of his attack we get a collection of predicates applied to the pure Ego which serve as a wholesome corrective to some of the proud names heaped upon it before. The Ego, he says, is "a merely logical qualitative unity of self-consciousness in thought generally;" it is in itself a perfectly empty or contentless idea-a perfectly empty expression which I can apply to every thinking subjectnay, it is actually "the poorest of all our ideas." No doubt the argument here is overlaid in parts by extraneous considerations, and infected by Kant's relativistic prejudice; but in pointing

out the merely logical character of the self reached by the analysis of knowledge, he is not only guided by a sounder instinct, but shows also a keener insight than his speculative followers. The logical exposition of thought in general is mistaken," he says, "for a metaphysical determination of the object." The words are spoken of the metaphysical psychologists, but it would be impossible to characterise more aptly the fallacy which underlies the Neo-Kantian deification of the abstract unity of thought.

APPENDIX TO LECTURE I.

Though it is hardly, perhaps, an integral part of the present argument, it seems natural to connect Kant's refusal to substitute for the real self a purely logical or formal unity with his refusal to identify the reality of the external world with mere relations. Kant's doctrine of things-in-themselves, as ordinarily understood, I cannot but hold to be fundamentally false, and a fruitful source of error;1 but it does not therefore follow that the whole external world is nothing more than a complex of thought-relations. There seems no reason why, if we resolve the rest of the external world in this way, we should not reduce our fellow-men also to mere complexes of relations, which have no existence on their own account. For our fellow-men are given to us, in the first instance, as part of the external world; and it would seem as if the same reasons which make us assign to them an existence on their own account, and not as mere objects either of our own or of a supposed universal consciousness, should lead us to attribute an (at least analogously) independent existence to the external world, or at any rate to certain existences in it. Kant himself, after the promulgation of his Critical system, was resolutely averse to speculation beyond certain limits; but there are indications in his writings that, if indulged, his speculations would have led him in a Leibnitian direction, as was indeed natural in the case of one who had been reared and

1 The fifth lecture of the previous course was chiefly devoted to combating the doctrine of the unknowable thing per se, as it appears in Kant and Hamilton.